Table 1.

Viral traits & features used to build our mammalian models.

Category	Viral Feature	Data type	Reason for inclusion
Host-driven	Mean phylogenetic distance between hosts	Continuous	Capturing phylogenetic and ecological distances between each virus’ known hosts and each mammal in our study.
	Mean ecological distance between hosts
	Maximum phylogenetic breadth4		Greater phylogenetic breadth indicates more generalist potential of the virus.
Virus genome & capsid	RNA	Binary	RNA viruses mutate/adapt faster65, and are generally deactivate quickly when exposed to the environment.
	Retro-transcribing		Retroviruses are generally very conserved66, have to enter the nucleus67 and insert into the genome. Additional steps may require specificity and limit range.
	Negative sense/positive sense		Sense affects replication cycle and range of host enzymes needed.
	Circular/linear		Circular/linear genome affects enlisting host enzymes for replication and translation68.
	Monopartite/segmented		Segmented viruses can undergo recombination if two strains of the same virus infect a cell69. This can lead to host range changes of segments of the genome.
	Enveloped		Envelopes are derived from the host cell membrane, so can affect specific-host immune activation. Enveloped viruses deactivate rapidly in the external environment (often requiring direct transfer). The envelope will change upon infection of a new host70.
	GC-content	Continuous	High GC content usually leads to higher thermo-stability of the genome71.
	Genome size		Genome size is indicative of many aspects of the virus such as complexity, DNA/RNA, and replication type. Replication site is linked to RNA/DNA genome – if a virus has a DNA stage it must replicate in the nucleus and overcome additional cell barriers.
Virus replication, release, and cell entry	Cytoplasm	Binary
	Release	Categorical	Affects rate of virus production, cell life-span and means of presentation to the immune system72.
	Cell entry		Availability of receptors influences potential host range.
Transmission routes	8 main transmission routes	Binary for each route	Route(s) of transmission affected by structure/stability of virus and nature of interaction between potential hosts.

We trained a suite of models for each mammalian species with two or more known viruses (n = 699). Each model comprised the below described features (response variable = 1 if the virus is known to associate with the focal mammalian species, 0 otherwise – methods section provides further details). Full description of these features, their sources and justification are listed in Supplementary Note 2.